In article <hersh.715458770@xenon.lcs.mit.edu>, hersh@expo.lcs.mit.edu (Jay Hersh) writes:
|>
|> I don't follow your conclusion at all Bill. All this fellow said was that
|> there is a source of free PEX Servers. He made no connection between costs
|> of commercial PEX Servers and the fact that they're based on a free (no quotes)
|> SI. Yes in fact the vendors paid for the SI, but if they didn't pay for the
|> SI they would all have had to pay for individual implementations themselves.
|> Since most vendors leverage some or all of the SI code, they do acheive some
|> savings in their development costs. I would guess the same is true for OpenGL,
|> but it is not accurate or fair to compare the development cost of PEX to the
|> license cost of OpenGL. The cost of the PEX-SI now (and when 6.0 is available)
|> to the general public will continue to be $0.00, it is publicly ftp-able
|> software. For many folks this is the bottom line they're concerned with,
|> how much can I get it for, not who paid to develop it, or how much it cost
|> them..
Jay, my disagreement with the original post was with the linkage between
the fact the the PEX-SI is ftp'able at no charge and the unstated assertion
that PEX would be therefore 'cheaper' than OpenGL.
For the vast majority of the computing world, a 3D API will be acquired
via a hardware purchase (an SGI or HP workstation) or a system software
purchase (Microsoft NT, Solaris, Destiny). The price these hardware
and software vendors will charge for it reflects their total development
costs.
Development costs equal acquisition costs (technology licensing fees, source
code licensing fees) AND implementation costs (engineers
to complete the product, create the documentation, and support the product).
I argue that the implementation costs are PEX are greater than for OpenGL, because of the existence of a more complete OpenGL SI and the Conformance Test.
Others on the net have concurred with this opinion.
Any knowledgeable economist (or adept finance manager) should tell you that
costs drives pricing decisions. And price means, in your terms, 'how much I
can get it for'. So, if the total costs of development are roughly equal (for
argument's sake only), the price to the end user for both technologies should
be the same, especially since they purport to offer much of the same
functionality to end users. Now, I accept that some firms may choose to act
economically irrational for strategic reasons (a workstation vendor may give PEX
or OpenGL software away for free to sell hardware). But since there is vendor
and ISV support for both PEX and OpenGL, I don't see any reason why one
technology would be the prefered product to 'subsidize'.
It is entirely appropriate and in fact imperative to compare the implementation
cost of PEX and the licensing cost of OpenGL, since they are both integral
parts of the total equation used in coming up with a price. My original point
was the total costs of developing OpenGL are no greater than PEX. Therefore
one should not assume that PEX will be cheaper than OpenGL.
For the Unix hacker needing to screw around with source code, the fact
that PEX can be ftp'ed for no charge is a good thing. But a mainstream
market is not driven by end users who need to have free access to source code.
No one complains in the Mac market that Quickdraw isn't freely available in source form. Same with GDI and Windows. In fact, the majority of X Window
users probably won't want source code. They just want applications that run
under X. Many of the places that want source are universities. OpenGL
source is available for $500 to them.
The argument that 'PEX is free' is terribly misleading. It clouds the real
issues in the discussion.
A simple analogy in the car business might prove my point:
OptionA: The car itself is free, but you need to overhaul the engine,
install a new exhaust system, and add new tires to get
to run. Once it's up and running, you have to buy premium gas for it, and
you only get 10 miles per gallon.
Option B: You must purchase the car for X dollars. Only a tuneup is required.
The car runs on unleaded regular, and is inexpensive to insure.
It is unclear from the description which option is cheaper in the long term.
Just because the car itself is free is no clear indicator of the total costs of
ownership.
What will drive the car purchase decision, assuming the costs are equal, are
other factors: performance, styling, status, roominess, the decibel level of the
stereo system..... It depends on your particular needs. There is no moral
superiority to Option A, except to people who just like to mess around with
